Add code to the loader to enforce the invariant that there is only a
single level of 'outer' symbol nesting. That is, if outer(X) = Y, then
outer(Y) is always zero.
Revise foldSubSymbolOffset based on the new invariant, allowing it to
be inlined, and then fix the various "for s.Outer != nil" loops in the
linker to just use an "if" instead of a loop.
Change-Id: Ib895702bc6de52718248f09a5368b84cb2e0a3fa
Reviewed-on: https://go-review.googlesource.com/c/go/+/231137
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Now that the loader's internal storage mechanism for symbol alignment
is array-based and not map-based, we can go back to computing symbol
alignment in the parallel-by-section section of dodata.
With this patch plus the previous one, this produces a small
kubelet speedup:
$ benchstat out.devlink.txt out.align.txt
name old time/op new time/op delta
RelinkKubelet 13.3s ± 2% 13.1s ± 2% -1.27% (p=0.000 n=20+20)
RelinkKubelet-WithoutDebug 7.36s ± 5% 7.14s ± 3% -3.00% (p=0.000 n=20+20)
Change-Id: I9eb0e8fea6aeb12f188f499e9031d5a3a23232c7
Reviewed-on: https://go-review.googlesource.com/c/go/+/231221
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Currently, for external relocations, the ExtReloc structure
contains all the fields of the relocation. In fact, many of the
fields are the same with the original relocation. So, instead, we
can just use an index to reference the original relocation and
not expand the fields.
There is one place where we modify relocation type: changing
R_DWARFSECTREF to R_ADDR. Get away with it by changing
downstreams.
It also makes it easier to retrieve the reloc variant.
This reduces some allocation. Linking cmd/compile with external
linking,
name old alloc/op new alloc/op delta
Reloc_GC 34.1MB ± 0% 22.7MB ± 0% -33.30% (p=0.000 n=5+4)
Change-Id: Id08a89ed2aee705296886d3b95014b806a0d55cf
Reviewed-on: https://go-review.googlesource.com/c/go/+/231217
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Revise the signature for "relocsym" to reflect the fact that many of
its arguments are invariant: push the invariant args into a struct and
pass the struct by reference.
Add a facility for doing batch allocation of external relocations in
relocsym, so that we don't wind up with wasted space due to the
default "append" behavior.
This produces a small speedup in linking kubelet:
$ benchstat out.devlink.txt out.dodata.txt
name old time/op new time/op delta
RelinkKubelet 14.2s ± 2% 13.8s ± 2% -3.11% (p=0.000 n=19+19)
RelinkKubelet-WithoutDebug 8.02s ± 3% 7.73s ± 3% -3.67% (p=0.000 n=20+20)
Change-Id: I8bc94c366ae792a5b0f23697b8e0108443a7a748
Reviewed-on: https://go-review.googlesource.com/c/go/+/231138
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Tweak doDataSect to reduce symbol sorting overhead, and calculate size
ahead of allocating the ctxt.datap slice. Yields a small speedup
(2-3%) linking kubelet.
Change-Id: I82869f5276caa4bee9f6e6f41da2b240e601ce50
Reviewed-on: https://go-review.googlesource.com/c/go/+/231047
Run-TryBot: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
We call (or will call) ResolveABIAlias in many places. Doing zero
symbol check everytime is annoying. Fold the condition into
ResolveABIAlias.
Change-Id: I10485fe83b9cce2d19b6bd17dc42176f72dae48b
Reviewed-on: https://go-review.googlesource.com/c/go/+/231046
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Adddynrel2 is a function pointer. In dynrelocsym we pass &r to
it, which will cause r to escape. Pass it by value instead.
Linking cmd/compile,
name old alloc/op new alloc/op delta
Dodata_GC 15.8MB ± 0% 5.9MB ± 0% -62.55% (p=0.008 n=5+5)
Change-Id: Ib86005d1026ebaca57777b27ead037e613585f44
Reviewed-on: https://go-review.googlesource.com/c/go/+/231045
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Archreloc2 is a function pointer. It will escape its pointer
arguments. In relocsym, as we pass &r and &rr to Archreloc2, it
causes them to escape, even if Archreloc2 is not actually called.
Instead, pass r by value. loader.Reloc2 is a small structure
which is intended to be passed by value.
For rr, as Archreloc2 will likely return true, we speculatively
add it to extRelocs slice and use that space to pass to
Archreloc2.
Linking cmd/compile,
name old alloc/op new alloc/op delta
Dwarfcompress_GC 110MB ± 0% 24MB ± 0% -78.34% (p=0.008 n=5+5)
Reloc_GC 24.6MB ± 0% 0.0MB ± 0% -100.00% (p=0.029 n=4+4)
Linking cmd/compile using external linking
name old alloc/op new alloc/op delta
Reloc_GC 152MB ± 0% 36MB ± 0% -76.07% (p=0.008 n=5+5)
Change-Id: I1415479e0c17ea9787f9a62453dce00ad9ea792f
Reviewed-on: https://go-review.googlesource.com/c/go/+/231077
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
And remove "2" from some function names.
Change-Id: Ibf1089970d849a42f53976064ceb9ade20bf6eba
Reviewed-on: https://go-review.googlesource.com/c/go/+/231017
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Non functional, but required by the majority of the architectures.
Change-Id: I57601016c28ce665a9d434e283a1db8bded9b133
Reviewed-on: https://go-review.googlesource.com/c/go/+/230858
Run-TryBot: Jeremy Faller <jeremy@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
When applying relocations, we need to resolve ABI aliases.
relocsym does that. Architecture-specific archreloc also needs to
do that. The old code doesn't do that since ABI aliases are
resolved in loadlibfull, or, in the old linker, in a much earlier
stage. We don't do this in the new linker, as we want to avoid
mutating relocations.
While here, move R_CONST and R_GOTOFF handling to generic code.
They appear on several architectures and the handling are same.
Should fix 386-clang and *bsd-386 builds.
Change-Id: I6681c94f0327555d6cf329d0a518c88848773671
Reviewed-on: https://go-review.googlesource.com/c/go/+/230857
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Support external linking for the new reloc pass as well, and
enable it on AMD64 and 386.
Change-Id: Ia71aec3d7c14e9d661e0748d2e988f29f220d1e1
Reviewed-on: https://go-review.googlesource.com/c/go/+/230308
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Only enabled for AMD64 when internal linking for now.
Change-Id: I2aa9ee47c0f7413ea7bbcdd31b8317c14220bba3
Reviewed-on: https://go-review.googlesource.com/c/go/+/230302
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
This change is just the mechanical work of moving the wavefront past
address.
Change-Id: I519ec49fa8ba50760c7d23fc084fcd3bb0544546
Reviewed-on: https://go-review.googlesource.com/c/go/+/229700
Run-TryBot: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The ReadOnly attribute was used to do copy on write when applying
relocations to symbols with read-only backing stores. Now that we
always apply relocations in the output buffer (mmap or heap), it
is always writeable. No need to tamper with the ReadOnly
attribute anymore.
Wasm is an exception, where we don't copy symbol contents to the
output buffer first. Do copy-on-write there.
This is in preparation of converting reloc to using the loader.
Change-Id: I15e53b7c162b9124e6689dfd8eb45cbe2ffd7153
Reviewed-on: https://go-review.googlesource.com/c/go/+/229991
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
It is supposed to work around symbol movement in machosymorder.
But machosymorder doesn't actually move symbols around.
Change-Id: Ibdc2ad41aaa8cd49e865088aa1ddb7ab399736cd
Reviewed-on: https://go-review.googlesource.com/c/go/+/230279
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
The symbol alignment is set based on its size. In dynreloc2
symbol size may change (e.g. elfdynhash2). So the alignment must
be set after dynreloc2.
Noticed this while debugging nondeterministic build on Solaris.
Idx Name Size VMA LMA File off Algn
8 .hash 000000c8 000000000048add2 000000000048add2 0008add2 2**3
CONTENTS, ALLOC, LOAD, READONLY, DATA
This doesn't look right, as the section address is not a multiple
of its alignment.
Change-Id: I23534cbc59695b7bc241838173fcc71dde95b195
Reviewed-on: https://go-review.googlesource.com/c/go/+/230278
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
This probably breaks darwin/arm64. Will fix.
Change-Id: I8be168985124f971e9d8ab5bc95c303336dd705b
Reviewed-on: https://go-review.googlesource.com/c/go/+/230019
Reviewed-by: Than McIntosh <thanm@google.com>
It should check the name of the symbol being added, not the
GC data symbol we're generating.
Change-Id: I123679778ee542b8d1f5c15bf090fa3578025c19
Reviewed-on: https://go-review.googlesource.com/c/go/+/230018
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Tweak the code in the amd64 version of adddynrel to avoid creating a
symbol updated for the symbol being processed until it's clear we need
to alter its relocations. This should help performance for the
PIE+internal linking scenario.
Reviewed-on: https://go-review.googlesource.com/c/go/+/229866
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Change-Id: Id25adfd81a5bbd2dde0f80a83b976397ba6abfb5
Reviewed-on: https://go-review.googlesource.com/c/go/+/230026
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
It doesn't do what it says. It has been like that since Go 1.4.
The current ouput is pretty useless. Remove it.
Change-Id: Id9b4ba04139aaf7ea59acbd51428b1c992115389
Reviewed-on: https://go-review.googlesource.com/c/go/+/229859
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
This patch begins the work of converting the linker's dodata phase to
work with loader APIs. Passes all.bash on linux/amd64, but hasn't been
tested on anything else (more arch-specific code needs to be written).
Use of the new dodata() phase is currently gated by a temporary
command line flag ("-newdodata"), and there is code in the linker's
main routine to insure that we only use the new version for the right
GOOS/GOARCH (currently restricted to ELF + AMD64).
Change-Id: Ied3966677d2a450bc3e0990e0f519b3fceaab806
Reviewed-on: https://go-review.googlesource.com/c/go/+/229706
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Create a new version of the GCProg type + methods that use loader APIs
instead of sym.Symbol.
This code isn't actually used just yet, but will be needed once the
wavefront reaches dodata() and we need to convert that phase.
Change-Id: I087521832015818204fe5c2ac99c7bd3f61b2bf0
Reviewed-on: https://go-review.googlesource.com/c/go/+/229037
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Currently, we use a dense array to store symbol's sections. The
array element is a *sym.Section, which takes 8 bytes per symbol
on a 64-bit machine. And the array is created upfront.
To reduce memory usage, use a 16-bit index for sections, so we
store 2 bytes per symbol. The array is pointerless, reducing GC
work. Also create the array lazily.
This reduces some memory usage: linking cmd/compile,
name old alloc/op new alloc/op delta
Loadlib_GC 42.1MB ± 0% 36.2MB ± 0% -14.01% (p=0.008 n=5+5)
name old live-B new live-B delta
Loadlib_GC 16.8M ± 0% 15.4M ± 0% -8.36% (p=0.008 n=5+5)
Archive_GC 98.2M ± 0% 97.2M ± 0% -1.02% (p=0.008 n=5+5) # at the end
Change-Id: If8c41eded8859660bca648c5e6fdf5830810fbf6
Reviewed-on: https://go-review.googlesource.com/c/go/+/229306
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Split out DWARF symbol-to-section assignment into its own separate
helper routine, to improve readability. No change in functionality.
Change-Id: Ic2e4f4d99afbff65161cbb8bd63e866ea555f322
Reviewed-on: https://go-review.googlesource.com/c/go/+/228957
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Additional cleanups and refactorings in the allocateSections portion
of dodata. Introduce some new helper routines to be used for common
cases in creating sections and assigning symbols, with a goal of
reducing duplicated code blocks and having more readable code.
No change in functionality.
Change-Id: I1b020b3ee993674329b2bebfd7c35995e3a2c043
Reviewed-on: https://go-review.googlesource.com/c/go/+/228883
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
This is more or less a direct translation, to get things going.
There are more things we can do to make it better, especially on
the handling of container symbols.
Change-Id: I11a0087e402be8d42b9d06869385ead531755272
Reviewed-on: https://go-review.googlesource.com/c/go/+/229125
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Create a couple of helper routines to aid in assigning symbols to
sections in dodata's allocateSections, then replace loops over symbol
lists with calls to the helpers, to reduce the amount of duplicate
code.
This patch also decouples gcprog/gcdata generation from
symbol-to-section assignment (previously intertwined), as an aid to
making the code less complicated.
No change in functionality.
Change-Id: If126579486bce458f697e32bad556df453df53e9
Reviewed-on: https://go-review.googlesource.com/c/go/+/228781
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Change linker DWARF generation to move away from emitting a single
giant list of DWARF symbols, and instead emit a list of descriptors,
with each descriptor holding the symbols for a specific DWARF section.
While placing all DWARF symbols in a single lists does come in handy
in certain instances, it also creates a lot of confusion and weird
code in other cases, specifically where we want to perform operations
on a section-by-section basis (resulting in code that tries to
re-discover section boundaries by walking/inspecting the list).
Change-Id: I4dac81bd38cba903c9fd7004d613597e76dfb77a
Reviewed-on: https://go-review.googlesource.com/c/go/+/228780
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Move more items into the dodata state object (including the "datsize"
variable used in allocateSections) and the Link ctxt pointer), so as
to prepare for follow-on refactorings. No change in functionality.
Change-Id: Ie2b1651c1ac9b89deb3f7692227dcd931240afa9
Reviewed-on: https://go-review.googlesource.com/c/go/+/228779
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Recreation of CL 228317.
The problem with that original CL was a late requested change,
reordering reloc and asmb, resulting in symbols having stale pointers to
their data. I've fixed this by preallocating the heap variable in OutBuf
for platforms w/o mmap.
Change-Id: Icdb392ac2c8d6518830f4c84cf422e78b8ab68c7
Reviewed-on: https://go-review.googlesource.com/c/go/+/228782
Run-TryBot: Jeremy Faller <jeremy@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Begin refactoring dodata to try to convert it from a single giant blob
to something more hierarchical, with descriptive function names for
sub-parts.
Add a state object to hold things like "data" and "dataMaxAlign"
arrays that are used throughout dodata. Extract out the code that
allocates data symbols to sections into a separate method (this
method is still too big, probably needs to be refactored again).
No change in functionality.
Change-Id: I7b52dc2aff0356e7d4b5d6f629d907fd37d3082c
Reviewed-on: https://go-review.googlesource.com/c/go/+/228259
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Hoist dwarfGenerateDebugSyms call up out of dodata to before
loadlibfull. This required a couple of small tweaks to the
loader and to loadlibfull.
Change-Id: I48ffb450d2e48b9e55775b73a6debcd27dbb7b9c
Reviewed-on: https://go-review.googlesource.com/c/go/+/228221
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
lib.Textp was used for text address assignment and trampoline
insertion. Now that it has been converted to using the loader,
no need to populate lib.Textp.
Port the logic of canonicalizing dupok symbol's package to the
loader.
unit.Textp was used for DWARF generation, which has also been
converted to using the loader.
Change-Id: I22d4dd30a52a29dd5b1b7b795d43a19f6215e4ac
Reviewed-on: https://go-review.googlesource.com/c/go/+/228140
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
We split the output into blocks and write them in parallel. The
block boundary is placed at symbol boundary. In the case of outer
symbols and sub symbols, currently we may split an outer symbol
into two blocks. This will be bad, as the two blocks will have
overlapping address range, since outer symbol and its sub symbols
occupies the same address range.
Make sure we place block boundary only at top-level symbol
boundaries.
Fix boringcrypto build.
Change-Id: I56811d3969c65c6be97672d8e1f1ea36b2447465
Reviewed-on: https://go-review.googlesource.com/c/go/+/227957
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Don't emit an error for undefined symbol in trampoline
generation pass, which will be duplicate as we'll emit a better
one later.
Fix TestUndefinedRelocErrors on PPC64.
Change-Id: I964d4bd63ec4b4c6eb5d98caf68db93ce6488bf5
Reviewed-on: https://go-review.googlesource.com/c/go/+/227617
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Implement text address assignment and trampoline generation using
the loader.
Note: the trampoline insertion part doesn't actually work. It
also needs to propagate Aux symbols for external symbols in
LoadFull. But it won't be needed after converting pclntab
generation, so I'll leave it out for now. This could break
linking large binaries on PPC64 and ARM.
Change-Id: Ie46a35b25d7c027983dd877207cfa8f67c32530b
Reviewed-on: https://go-review.googlesource.com/c/go/+/227482
Reviewed-by: Than McIntosh <thanm@google.com>
This ports CL 226997 to the dev.link branch.
- The assembler part and old object file writing are unchanged.
- Changes to cmd/link are applied to cmd/oldlink.
- Add alignment field to new object files for the new linker.
Change-Id: Id00f323ae5bdd86b2709a702ee28bcaa9ba962f8
Reviewed-on: https://go-review.googlesource.com/c/go/+/227025
Reviewed-by: Than McIntosh <thanm@google.com>
Make sure we never overrun the end address.
Should fix AIX build.
Change-Id: I9926387d4512ec8b4acc32b7f32cee2b2aca25b4
Reviewed-on: https://go-review.googlesource.com/c/go/+/226718
Reviewed-by: Jeremy Faller <jeremy@golang.org>
